os/doc
3
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
doc/en/smp/index.sgml
Simon L. B. Nielsen 7fe5c10c82 MFP4: 
- Merge more of the new layout.

Submitted by:	Emily Boyd <emilyboyd@emilyboyd.com>
Sponsored by:	Google Summer of Code 2005
2005-10-04 22:27:39 +00:00

2647 lines
69 KiB
Text
Raw Blame History

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" [
<!ENTITY base CDATA "..">
<!ENTITY date "$FreeBSD: www/en/smp/index.sgml,v 1.155 2005/10/04 19:43:53 hrs Exp $">
<!ENTITY title "FreeBSD SMP Project">
<!ENTITY email 'freebsd-smp'>
<!ENTITY % navincludes SYSTEM "../includes.navdevelopers.sgml"> %navincludes;
<!ENTITY % includes SYSTEM "../includes.sgml"> %includes;
<!-- Status levels -->
<!ENTITY status.done "<font color=green>Done</font>">
<!ENTITY status.wip "<font color=blue>In progress</font>">
<!ENTITY status.stalled "<font color=purple>Stalled</font>">
<!ENTITY status.new "<font color=red>Not Started</font>">
<!ENTITY status.resolved "<font color=green>Resolved</font>">
<!ENTITY status.unresolved "<font color=red>Unresolved</font>">
<!-- The list of contributors was moved to a separate file so that it can
be used by other documents in the FreeBSD web site. -->
<!ENTITY % developers SYSTEM "../developers.sgml"> %developers;
]>
<html>
&header;
<h2>Contents</h2>
<ul>
<li><a href="#goal">Project Goal</a></li>
<li><a href="#plan">Project Plan</a></li>
<li><a href="#resources">Resources and Links</a></li>
<li><a href="#subsystems">Subsystems and Staffing</a></li>
<li><a href="#status">Tasks</a></li>
<li><a href="#issues">Known Issues</a></li>
<li><a href="#news">News</a></li>
</ul>
<a name="goal"></a>
<h2>Project Goal</h2>
<p>The FreeBSD Symmetric MultiProcessing (SMP) project, often referred to
as SMPng (SMP next generation), is focused on introducing parallelism
into the FreeBSD kernel.
While earlier versions of FreeBSD (3.x, 4.x) supported parallelism in
user processes, the kernel was limited to executing on a single
processor at a time, using what is referred to as a "Giant lock" around
the kernel.
For many interesting workloads, this results in a substantial speed-up,
as significant computation occurs in user processes, especially for
applications such as rendering and compilation.
However, for kernel-intensive applications, such as intensive network
or file system I/O, contention on the kernel lock results in little
speed-up.
The end goal of the SMPng Project is to decompose the Giant lock into a
number of smaller locks, resulting in reduced contention (and improved
SMP performance).
However, important steps along the way include redesigning significant
portions of the FreeBSD kernel architecture around the notion of
ubiquitous parallelism: that at any moment, many processors might enter
the kernel at the same time.
This includes the introduction of more mature threading and
synchronization primitives, interrupt threads, cache-aware allocation
and scheduling, and topology-aware scheduling. </p>
<p>SMPng debuted in FreeBSD 5.0-RELEASE in January of 2003, and involved
over five years of continuous development by a large number of members
of the FreeBSD development team, as well as many external contributors.
Since 5.0, the implementation has substantially matured; in the 5.0
release, the basic architectural changes required to support SMPng were
complete, including new kernel memory allocators, synchronization
routines, the move to ithreads, and the removal of the Giant lock from
activities such as process scheduling and several common forms of IPC.
Since 5.0, the architecture has been refined in a number of ways,
including optimizing synchronization approaches, stability and
performance testing on larger systems (up to 12 processors), and the
removal of Giant from several significant parts of the operating
system, including large parts of Virtual Memory (VM), the Virtual File
System (VFS), UNIX File System (UFS), most parts of the network stack,
including IPv4, IPV6, FAST_IPSEC, UNIX domain sockets, and NetGraph,
and also from additional inter-process communication primitives.
The SMP-aware kernel slab allocator is now used almost universally, and
the focus has changed from "make it work", to "optimize it!".
Simultaneous work on KSE 1:1 and M:N threading has also allowed
applications to take advantage of new kernel parallelism.
The FreeBSD 5.3 kernel introduced Giant-free network stack execution
for most relevant code paths, and the FreeBSD 6.x kernel introduces
MPSAFE VFS, as well as wide-spread performance optimization.</p>
<p>Continuing work on SMPng includes work sweeping up the "loose ends"
that remain under Giant, such as parts of NFS, less widely used file
systems such as NTFS, and less commonly used network stack components,
such as SLIP.
Another important focus is performance measurement and optimization,
which build on and refine the SMPng architecture: the introduction of
features such as the kernel trace facility (KTR), hardware performance
monitor counters (hwpmc), lock profiling, and improved memory
monitoring play important roles in this process.
Other important debugging and testing facilities include WITNESS, a
run-time kernel lock order verifier, and wide-spread use of lock
assertions and run-time invariants testing.</p>
<p>The FreeBSD Project recognizes, in particular, the contributions of
the BSD/OS development team, including architectural direction and
suggestions, as well as for initial locking strategies for some system
components, and source code for some of the starting primitives, not to
mention BSDI's contributions in staffing and resources for early parts
of this project.</p>
<a name="plan"></a>
<h2>Project Plan</h2>
<p>This web page contains information relating to the SMPng effort;
because of the large amount of work and rapid pace of development,
it can fall a bit behind reality.</p>
<p>The task list below is not intended to be complete, but does
represent a set of relevant and/or important components of the
overall work. The "Responsible" field identifies a developer who
has expressed willingness to be responsible for completing the
identified task; this doesn't preclude others working on it,
but suggests that coordination with the responsible party might
be appropriate so as to avoid unnecessary duplication of work,
and to maximize forward progress. If beginning work on a new
area of substantial size, or one that appears unclaimed, it may
be worth dropping an e-mail to the FreeBSD SMP mailing list to
see if any progress has been made.</p>
<p>The definition of the date field varies depending on the status
of a task. For completed tasks, it refers to the date completed
or reported completed. For in-progress tasks, it refers to the
date of the last update of the entry. For stalled tasks, it
refers to the date that the task was declared stalled. For
new tasks, it refers to the date the task was added to the list.</p>
<p>Locking down of individual device drivers is tracked at
<a href="http://www.FreeBSD.org/projects/busdma/">the busdma and
SMPng driver conversion webpage</a>.</p>
<p>Network stack locking information is available at Robert Watson's
<a href="http://www.watson.org/~robert/freebsd/netperf/">netperf web
page</a> and the FreeBSD.org <a href="../projects/netperf/">Netperf
web page</a>.
An SMP <a href="../projects/netperf/cluster.html">network performance
cluster</a> has also been created for the purposes of testing.</p>
<p>Tasks are sorted first by status, then by date.</p>
<a name="resources"></a>
<h2>Resources and Links</h2>
<ul>
<li><p>Most SMP-related discussion takes place on the
<a href="mailto:freebsd-smp@FreeBSD.org">freebsd-smp</a>
mailing list. You can read more about mailing lists in the
<a href="../doc/en_US.ISO8859-1/books/handbook/eresources.html#ERESOURCES-MAIL">
Electronic Resources chapter</a> of the
<a href="../doc/en_US.ISO8859-1/books/handbook/index.html">FreeBSD Handbook</a>.</p></li>
<li><p>Steve Passe has been maintaining a
<a href="http://people.FreeBSD.org/~fsmp/SMP/SMP.html">SMP
project page</a> that contains additional information, and
goes back further in time than this web page.</p></li>
<li><p>&a.rwatson; is maintaining a page for SMP-related network
performance work for the <a href="../projects/netperf/">Netperf
project</a>. In addition, he has a <a
href="http://www.watson.org/~robert/freebsd/netperf/">a personal
web page with a change log and other information</a>.</p></li>
<li><p><a href="http://www.osnews.com/">OSNews</a> has run an
interview with FreeBSD developers &a.rwatson;, &a.jhb;, and
&a.scottl;, who talk about many features in 6.x, including recent
SMPng work, SMPVFS, and more. The <a
href="http://www.osnews.com/story.php?news_id=10951">article is
here.</a></p></li>
<li><p>A series of man pages on kernel synchronization and threading
primitives can be found here:</p>
<ul>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=mutex&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Mutexes (mutex(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=mtx_pool&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Mutex Pools (mtx_pool(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=critical&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Critical Sections (critical(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=sx&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Shared/Exclusive Locks (sx(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=condvar&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Condition Variables (condvar(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=sema&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Semaphores (sema(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=sleep&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel sleep/wakeup API (sleep(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=kthread&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Kernel Threads (kthread(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=ithread&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Interrupt Threads (ithread(9))</a></li>
<li><a href="http://www.freebsd.org/cgi/man.cgi?query=swi&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">Software Interrupt Threads (swi(9))</a></li>
</ul>
<p>In addition, the old <a href="http://www.freebsd.org/cgi/man.cgi?query=spl&amp;apropos=0&amp;sektion=9&amp;manpath=FreeBSD+6.0-current">SPL man page</a> may be of interest, as it reflects the previous SMP synchronization model.</p>
</li>
<li><p>
Hiten Pandya's <a
href="http://storm.uk.FreeBSD.org/~hiten/smp_synch_rules.html">
SMP kernel synchronization rules</a>.</p></li>
<li><p>&a.dillon; is maintaining a
<a href="http://apollo.backplane.com/FreeBSDSmp/">web page</a>
that documents the work he is doing on the SMP project.</p></li>
<li>
<p>
<a href="http://www.usenix.org/events/bsdcon02/full_papers/baldwin/baldwin_html/index.html">
"Locking in the Multithreaded FreeBSD Kernel" USENIX paper by &a.jhb;</a>.
</p>
</li>
<li>
<p>
<a href="http://www.usenix.org/publications/library/proceedings/bsdcon03/tech/hsu.html">
"Reasoning about SMP in FreeBSD" BSDCon'03 paper by &a.hsu;</a>.
</p>
</li>
<li>
<p>
<a href="http://www.usenix.org/publications/library/proceedings/bsdcon03/tech/roberson.html">
"ULE: A Modern Scheduler for FreeBSD" BSDCon'03 paper by &a.jeff;</a>.
</p>
</li>
</ul>
<a name="subsystems"></a>
<h2>Subsystems and Staffing</h2>
<p>This is an incomplete list of high-level kernel subsystems and
current, active staff working on SMP architecture and stability.</p>
<table class="tblbasic">
<tr class="heading">
<th> Subsystem </th>
<th> Status </th>
<th> Last updated </th>
<th> Staffing </th>
</tr>
<tr>
<td> Newbus </td>
<td> &status.wip; </td>
<td> 5 October 2003 </td>
<td> &a.imp;</td>
</tr>
<tr>
<td> VM </td>
<td> &status.wip; </td>
<td> 4 October 2003 </td>
<td> &a.alc;</td>
</tr>
<tr>
<td> Buffer cache </td>
<td> &status.wip; </td>
<td> 4 October 2003 </td>
<td> &a.jeff;, &a.phk; </td>
</tr>
<tr>
<td> VFS </td>
<td> &status.wip; </td>
<td> 4 October 2003</td>
<td> &a.jeff; </td>
</tr>
<tr>
<td> Processes and thread operations </td>
<td> &status.wip; </td>
<td> 5 May 2003 </td>
<td> &a.jhb; </td>
</tr>
<tr>
<td> Scheduler </td>
<td> &status.done; </td>
<td> 23 April 2004 </td>
<td> &a.jhb;, &a.jeff; </td>
</tr>
<tr>
<td> GEOM </td>
<td> &status.done; </td>
<td> 5 February 2003 </td>
<td> &a.phk; </td>
</tr>
<tr>
<td> File descriptors </td>
<td> &status.done; </td>
<td> 5 February 2003 </td>
<td> &a.alfred;, &a.tanimura; &a.rwatson; </td>
</tr>
<tr>
<td> TTY subsystem </td>
<td> &status.wip; </td>
<td> 24 July 2004 </td>
<td> &a.phk; </td>
</tr>
<tr>
<td> Pipe IPC </td>
<td> &status.done; </td>
<td> 4 October 2003 </td>
<td> &a.alfred; </td>
</tr>
<tr>
<td> Socket structures and system calls </td>
<td> &status.done; </td>
<td> 25 November 2004 </td>
<td> &a.sam;, &a.rwatson; </td>
</tr>
<tr>
<td> KQueue </td>
<td> &status.done; </td>
<td> 24 November 2004 </td>
<td> &a.jmg;, &a.green; </td>
</tr>
<tr>
<td> IPv4 </td>
<td> &status.done; </td>
<td> 23 April 2004 </td>
<td> &a.jennifer;, &a.hsu;, &a.sam;, &a.rwatson;, &a.gnn; </td>
</tr>
<tr>
<td> IPv6 </td>
<td> &status.wip; </td>
<td> 01 July 2005 </td>
<td> &a.rwatson;, &a.gnn; </td>
</tr>
<tr>
<td> IPX/SPX </td>
<td> &status.done; </td>
<td> 09 January 2005 </td>
<td> &a.rwatson; </td>
</tr>
<tr>
<td> netatalk </td>
<td> &status.done; </td>
<td> 02 February 2005 </td>
<td> &a.rwatson; </td>
</tr>
<tr>
<td> Network stack infrastructure </td>
<td> &status.wip; </td>
<td> 4 October 2003 </td>
<td> &a.hsu;, &a.sam;, &a.rwatson;, &a.mlaier;, &a.luigi;, Maurycy Pawlowski-Wieronski &lt;maurycy@fouk.org&gt;, &a.brooks;,
&a.rik; </td>
</tr>
<tr>
<td> NFS Client </td>
<td> &status.wip; </td>
<td> 23 April 2004 </td>
<td> &nbsp; </td>
</tr>
<tr>
<td> NFS Server </td>
<td> &status.wip; </td>
<td> 24 November 2004 </td>
<td> &a.rwatson; </td>
</tr>
</table>
<a name="status"></a>
<h2>Tasks</h2>
<p>Following is an incomplete list of general tasks.</p>
<table class="tblbasic">
<tr class="heading">
<th> Task </th>
<th> Responsible </th>
<th> Last updated </th>
<th> Status </th>
</tr>
<tr>
<td> Convert the giant lock from spinning to blocking, add the
scheduler lock, add per-CPU idle processes. </td>
<td> &a.dillon; </td>
<td> 25 June 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Port the BSD/OS locking primitives (i386). </td>
<td> &a.jake; </td>
<td> 3 July 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement heavy-weight interrupt threads (i386). </td>
<td> &a.grog; </td>
<td> 3 August 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Rewrite the low level interrupt code (i386 UP). </td>
<td> &a.grog; </td>
<td> 3 August 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Demonstrated reasonable stability (self-hosted buildworld)
(i386 UP). </td>
<td> -smp developers </td>
<td> 12 August 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Port the BSD/OS locking primitives (alpha). </td>
<td> &a.dfr; </td>
<td> 24 August 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Stub out (disable) spl()s. </td>
<td> &a.grog; </td>
<td> 30 August 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Port the BSD/OS ktr code. </td>
<td> &a.grog;, &a.jhb; </td>
<td> 30 August 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Rewrite the low level interrupt code (i386 SMP). </td>
<td> &a.jhb; </td>
<td> 1 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Demonstrated reasonable stability (self-hosted buildworld)
(i386 SMP). </td>
<td> -smp developers </td>
<td> 6 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Demonstrated reasonable stability (self-hosted buildworld)
(alpha). </td>
<td> -smp developers </td>
<td> 6 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make malloc and friends thread-safe. </td>
<td> &a.jasone; </td>
<td> 10 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement msleep(), make tsleep() an msleep() wrapper. </td>
<td> &a.jake; </td>
<td> 11 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make fxp driver thread-safe. </td>
<td> &a.cp; </td>
<td> 17 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make mbuf's thread-safe. </td>
<td> &a.bmilekic; </td>
<td> 29 September 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock manager re-work. </td>
<td> &a.jasone; </td>
<td> 3 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement heavy-weight interrupt threads (alpha). </td>
<td> &a.jhb;, &a.dfr; </td>
<td> 5 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Rewrite the low level interrupt code (alpha). </td>
<td> &a.dfr;, &a.jhb; </td>
<td> 5 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Process accounting. </td>
<td> &a.tegge;, &a.jhb; </td>
<td> 5 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make ethernet drivers thread-safe. </td>
<td> &a.wpaul; </td>
<td> 15 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make the mutex headers mostly machine-independent. </td>
<td> &a.jhb; </td>
<td> 20 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Rename SMP_DEBUG to MUTEX_DEBUG. </td>
<td> &a.jhb; </td>
<td> 20 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Give each soft interrupt its own thread. </td>
<td> &a.cp; </td>
<td> 25 October 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make sf_bufs (sendfile(2)) thread-safe. </td>
<td> &a.bmilekic; </td>
<td> 5 November 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make the witness code work correctly. </td>
<td> &a.jhb; </td>
<td> 18 November 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Split the ktr-specific code out of db_interface.c. </td>
<td> &a.jhb; </td>
<td> 15 December 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Convert the sio driver to using a spin mutex. </td>
<td> &a.jhb; </td>
<td> 18 December 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement condition variables. </td>
<td> &a.jake;, &a.jasone; </td>
<td> 15 January 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add a flag to mtx_init() (MTX_RECURSE) that denotes
whether a mutex is allowed to recurse. </td>
<td> &a.bmilekic; </td>
<td> 19 January 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make the zone allocator thread-safe. </td>
<td> &a.des; </td>
<td> 21 January 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Convert simplelocks to mutexes. </td>
<td> &a.jasone; </td>
<td> 24 January 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make kernel preemptive with respect to interrupts. </td>
<td> &a.jake; </td>
<td> 31 January 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Cleanup of mutex API. </td>
<td> &a.bmilekic; </td>
<td> 8 February 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Remove COM_LOCK. </td>
<td> &a.markm; </td>
<td> 11 February 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Merge various scheduling classes into one run queue.
Modify scheduler to support preemptable kernel. </td>
<td> &a.jake; </td>
<td> 11 February 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make priority propagation work correctly. </td>
<td> &a.jake; </td>
<td> 11 February 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make most of the interrupt thread code MI and shared
between hardware and software interrupts. </td>
<td> &a.jhb; </td>
<td> 18 February 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add protection to struct jail and jail-related functionality. </td>
<td> &a.rwatson; </td>
<td> 20 February 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement sx (shared/exclusive) locks. </td>
<td> &a.jasone; </td>
<td> 5 March 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Generalize/improve witness to handle more complex locking
primitives (mtx, sx). </td>
<td> &a.jhb; </td>
<td> 28 March 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Convert the allproc and proctree locks from lockmgr locks
to sx locks. </td>
<td> &a.jhb; </td>
<td> 28 March 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make mbuf system use condition variables instead of
msleep()/wakeup(). </td>
<td> &a.bmilekic; </td>
<td> 2 April 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Remove &lt;sys/mutex.h&gt; includes from other kernel
headers such as &lt;vm/vm_zone.h&gt;, &lt;sys/resourcevar.h&gt;,
&lt;sys/ucred.h&gt;, and &lt;sys/mbuf.h&gt;. </td>
<td> &a.markm; </td>
<td> 15 May 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Cleanup the various mp_machdep.c's, unify various SMP
API's such as IPI delivery, etc. </td>
<td> &a.jhb; </td>
<td> 15 May 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make most of the forward_* and forwarded_* functions MI. </td>
<td> &a.jhb; </td>
<td> 15 May 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Complete the MD support for SMP on the Alpha platform.</td>
<td> &a.gallatin;, &a.dfr;, &a.jhb; </td>
<td> 15 May 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Convert select() to use condition variables. </td>
<td> &a.tanimura; </td>
<td> 15 May 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add a "giant" lock around the VM subsystem. </td>
<td> &a.alfred; </td>
<td> 13 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Introduce a modified slab allocator for the mbuf subsystem. </td>
<td> &a.bmilekic; </td>
<td> 21 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add a witness_assert() function to handle lock assertions. </td>
<td> &a.jhb; </td>
<td> 27 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Extend sx locks to support try lock operations. </td>
<td> &a.jhb; </td>
<td> 27 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Document KTR. </td>
<td> &a.jhb; </td>
<td> 28 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make fork_return, fork_exit, ast, and userret MI. </td>
<td> &a.jhb; </td>
<td> 29 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make sched_lock's savecrit a per-process property saved
and restored in mi_switch and initialized in fork_exit. </td>
<td> &a.jhb; </td>
<td> 30 June 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make ast() loop. </td>
<td> &a.jhb; </td>
<td> 10 August 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add upgrade/downgrade sx lock operations. </td>
<td> Alexander Kabaev, &a.jasone; </td>
<td> 13 August 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement semaphores. </td>
<td> &a.jasone; </td>
<td> 14 August 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add support for upgrade/downgrades in witness. </td>
<td> &a.jhb; </td>
<td> 23 August 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make most of cpu_wait() and cpu_exit() MI. </td>
<td> &a.peter; </td>
<td> 9 September 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Split NFS into client and server. </td>
<td> &a.peter; </td>
<td> 18 Oct 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock taskqueues. </td>
<td> &a.arr;, &a.jhb; </td>
<td> 25 October 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add a per-thread ucred reference. </td>
<td> &a.jhb; </td>
<td> 25 October 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make most of the per-CPU stuff MI. </td>
<td> &a.jhb; </td>
<td> 11 December 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make critical section saved state per-thread instead of per-lock
so that interlocking spin locks work properly. </td>
<td> &a.jhb; </td>
<td> 17 December 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Replace the APIC-specific imen_mtx with a MI-named icu_lock to
protect interrupt controllers and associated data within the kernel
for both i386 and alpha. </td>
<td> &a.jhb; </td>
<td> 20 December 2001 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Use the per-thread critical section nesting level in the mutex
and interrupt thread code to automatically determine when to not
preempt. This makes the MTX_NOSWITCH, SWI_SWITCH, and
SWI_NOSWITCH flags obsolete as the kernel will be able to figure
out the proper behavior on its own. </td>
<td> &a.jhb; </td>
<td> 5 January 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock struct filedesc and struct file. </td>
<td> &a.tanimura;, &a.alfred; </td>
<td> 12 January 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock struct pgrp, struct session, and struct sigio. </td>
<td> &a.tanimura; </td>
<td> 23 February 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock pipe implementation, but not sigio/fown, VM interactions. </td>
<td> &a.alfred; </td>
<td> 27 February 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Move to explicit reference counting for soft vnode references.
</td>
<td> &a.phk; </td>
<td> 8 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Initialize mutex pools early enough that sx locks can be used
for VM. </td>
<td> &a.green; </td>
<td> 14 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Place a global lock (sellock) around selinfo structures to fix a
variety of lock order reversals, and make select() MP-safe. </td>
<td> &a.alfred;, &a.davidc; </td>
<td> 14 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Push down Giant on read, write, pread, pwrite system calls,
acquiring Giant in the per-subsystem fileop layer for sockets, VFS,
etc.</td>
<td> &a.alfred; </td>
<td> 15 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down kernel module structures. </td>
<td> &a.arr; </td>
<td> 18 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down kernel linker globals.</td>
<td> &a.arr; </td>
<td> 18 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Rewrite kernel memory allocator to be a slab allocator that
uses per-cpu caches. </td>
<td> &a.jeff; </td>
<td> 21 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Replace incorrect use of MD critical section API to disable
interrupts with a specific interrupt disable API. </td>
<td> &a.imp;, &a.dfr;, &a.benno;, &a.jhb; </td>
<td> 21 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down access to the shared p_args "process arguments"
structure through appropriate protection of that structure and
references to it. </td>
<td> &a.mini; </td>
<td> 31 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Move from flags/tsleep lock to sx locks to protect sysctl
tree from updates during sysctl operations. </td>
<td> &a.mini; </td>
<td> 1 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Create/port userland tool to manage KTR event dumps. </td>
<td> &a.jake; </td>
<td> 1 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Create MTX_SYSINIT and SX_SYSINIT macros that allow for
initializing locks that are subsystem independent. </td>
<td> &a.arr; </td>
<td> 2 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down the global securelevel variable. </td>
<td> &a.arr; </td>
<td> 2 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make grow_stack() MI. Possibly even a macro or inline. </td>
<td> &a.alc; </td>
<td> 6 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock use of p_fd, which otherwise can result in corrupted
p_fd panics during heavy operation. Start with a global,
and move to per-proc locking. </td>
<td> &a.alfred;, &a.tanimura; </td>
<td> 8 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock struct pargs. </td>
<td> &a.mini; </td>
<td> 9 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock sysctl hierarchy. </td>
<td> &a.mini; </td>
<td> 9 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make {o,}sigreturn() MPSAFE. </td>
<td> &a.alc; </td>
<td> 11 April 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Rewrite kernel memory allocator so that Giant is not required
for malloc() or free(). </td>
<td> &a.jeff; </td>
<td> 2 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Replace complex shared/exclusive locking scheme in the VM
system with a purely exclusive lockmgr locking scheme, simplifying
locking and removing potential livelock/deadlock scenarios. </td>
<td> &a.green;, &a.alc; </td>
<td> 3 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Push down Giant into readv/writev system calls in style of
read/write/pread/pwrite once malloc no longer requires Giant in
the handling of iovec structures for uio. </td>
<td> &a.alc; </td>
<td> 9 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Push down Giant in mprotect(), minherit(), and madvise() so
that it is no longer acquired and released directly. </td>
<td> &a.alc; </td>
<td> 18 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Update suser() and p_can*() APIs to accept threads instead of
processes. </td>
<td> &a.jhb; </td>
<td> 18 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Broadly transition to td_ucred from p_ucred once KSE dependencies
are in place. </td>
<td> &a.jhb; </td>
<td> 18 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add a witness_sleep() check to uma_zalloc() to catch code
calling malloc() or uma_zalloc() while holding non-sleepable
locks. </td>
<td> &a.jhb; </td>
<td> 20 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Optimize UP support by changing spin locks to only perform
critical section enter and exits. </td>
<td> &a.jhb; </td>
<td> 21 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make sleep mutexes spin if the current lock holder is
executing on another CPU. </td>
<td> &a.jhb; </td>
<td> 21 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add support for the IA32 pause instruction to spin loops in
locks. </td>
<td> &a.jhb; </td>
<td> 21 May 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make KTRACE write into tracefiles asynchronously. </td>
<td> &a.jhb; </td>
<td> 7 June 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Remove Giant from modnext(2), modfnext(2),
modstat(2),and modfind(2).</td>
<td> &a.arr; </td>
<td> 25 June 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Fix synchronization of TLB flushes and invlpg() on x86 SMP. </td>
<td> &a.peter; </td>
<td> 12 July 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing for mutex contention. </td>
<td> &a.iedowse; </td>
<td> 26 Augist 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make cpu_coredump MI. </td>
<td> &a.peter; </td>
<td> 7 September 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add a subsystem lock to the accounting code. </td>
<td> &a.arr; </td>
<td> 11 September 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Allow KTR(9) to write trace records to alq(9) record facility.
</td>
<td> &a.jeff; </td>
<td> 22 September 2002 </tD>
<td> &status.done; </td>
</tr>
<tr>
<td> Create mechanism in cdevsw structure to protect
thread-unsafe drivers. </td>
<td> &a.phk; </td>
<td> 27 September 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Fix SIGXPU and other #if 0'd things in mi_switch(). </td>
<td> &a.jhb; </td>
<td> 30 September 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down TrustedBSD MAC implementation. </td>
<td> &a.rwatson; </td>
<td> 11 November 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock eventhandlers. </td>
<td> &a.msmith;, &a.mini;, &a.jhb; </td>
<td> 11 March 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Fix PHOLD() so that it blocks to guarantee PS_INMEM. </td>
<td> &a.jhb; </td>
<td> 22 April 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Fix various procfs_machdep.c to not use sched_lock. </td>
<td> &a.jhb; </td>
<td> 22 April 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock all references to process credentials and remove Giant
from process credential-related system calls. </td>
<td> &a.jhb; </td>
<td> 1 May 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Merge the procsig and sigacts structures, move the new sigacts
structure out of the U-area and add appropriate locking. </td>
<td> &a.jhb; </td>
<td> 13 May 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Remove Giant from the kill() and killpg() system calls. </td>
<td> &a.jhb; </td>
<td> 13 May 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Enhance the mutex pool implementation to allow creation and
use of multiple, dynamically allocated pools with adjustable pool
sizes and mutex options. </td>
<td> &a.truckman; </td>
<td> 16 July 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Create mutex profiling tool for the kernel so as to measure
contention and behavior of kernel mutexes. </td>
<td> &a.eivind;, &a.des; </td>
<td> 31 March 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down linker_file_t structures in the kernel linker.</td>
<td> &a.arr; </td>
<td> 19 June 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock pipe implementation: VM optimizations. </td>
<td> &nbsp; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Reimplement i386 interrupt and SMP code so that SMP kernels
work on UP boxes and SMP can be enabled in GENERIC. </td>
<td> &a.jhb; </td>
<td> 3 November 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement generic turnstiles to use when blocking on non-sleepable
locks. </td>
<td> &a.jhb; </td>
<td> 11 November 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Split witness_lock() into witness_checkorder() and
witness_lock(). witness_checkorder() would be called before
acquiring a lock to increase the changes of detecting and warning
about a reversal prior to deadlocking. witness_lock() would
simply update witness' internal state to note that a lock has
been acquired. </td>
<td> &a.jhb; </td>
<td> 24 January 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock per-process resource limits. </td>
<td> &a.mtm;, &a.jhb; </td>
<td> 4 February 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Implement a sleep queue abstraction to be used by both msleep()
and condition variables. This new abstraction should use a hash
table of sleep queues with a spin lock on each sleep queue chain
similar to turnstile chain locks to make sched_lock finger
grained.</td>
<td> &a.jhb; </td>
<td> 27 February 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Remove Giant from jail(2). </td>
<td> &a.arr;, &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add subsystem locking to NFSv2, NFSv3 server, permitting upcalls
and other network-related elements to run Giant-free. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing for UMA allocation/free events. </td>
<td> &a.rwatson; </td>
<td> 05 August 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing for GEOM I/O events. </td>
<td> &a.rwatson; </td>
<td> 21 October 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing for busdma events. </td>
<td> &a.rwatson; </td>
<td> 23 October 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing for critical sections. </td>
<td> &a.rwatson; </td>
<td> 07 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make the kernel fully preemptive. </td>
<td> &a.jhb; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock pipe implementation: sigio/fown-related evil. </td>
<td> &a.alfred; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down the SysV IPC code. </td>
<td> &a.alfred; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock contention measurement tool to measure heat of various
locks, including Giant, and permit more directed performance and
locking strategy optimization. </td>
<td> &a.rwatson; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing to scheduler run queues. </td>
<td> &a.jeff; </td>
<td> 26 December 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Review locking strategy and correctness of VFS operations
and fix up various failure modes associated with enabling
VFS locking assertions. </td>
<td> &a.jeff; </td>
<td> 01 January 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Document in-vnode locking strategy, clean it up. </td>
<td> &a.jeff; </td>
<td> 01 January 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Run cross-file system VFS without Giant, acquiring Giant
conditionally based on a file system flag. </td>
<td> &a.jeff; </td>
<td> 01 January 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Run UFS file system MPSAFE. </td>
<td> &a.jeff; </td>
<td> 01 January 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) tracing for buffer cache events. </td>
<td> &a.jeff; </td>
<td> 24 January 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Break out critical section and spin lock APIs, and re-optimize
critical sections to not disable interrupts in hardware, due to
the high cost on some hardware architectures. </td>
<td> &a.jhb; </td>
<td> 04 April 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Modify uma(9) to use critical sections to protect per-CPU
statistics, instead of mutexdes, in order to optimize access. </td>
<td> &a.rwatson; </td>
<td> 29 April 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Migrate malloc(9) to per-CPU statistics, and use critical
sections to optimize access to those statistics. </td>
<td> &a.rwatson; </td>
<td> 29 May 2005 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add KTR(9) support for KTR_VFS to trace additional VFS events,
rather than mechanically inserted KTR_VOP events. </td>
<td> &a.jeff; </td>
<td> 11 June 2005 </td>
<td> &status.done; </td>
</tr>
<!-- WIP -->
<tr>
<td> Lock struct proc. </td>
<td> &a.jhb; </td>
<td> 20 February 2001 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock down the tty subsystem. </td>
<td> Dick Garner, Jeremy Scofield, &a.tmm;, &a.phk; </td>
<td> 24 July 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Fix clock locking to be the same on all platforms. </td>
<td> &a.jhb; </td>
<td> 16 November 2001 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Make use of process locking and process reference counting
to protect debugging interfaces (and procfs). </td>
<td> &a.jhb; </td>
<td> 27 February 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Make use of process locking to protect process monitoring
sysctls, including those employed by 'ps' and related tools.
<td> &a.jhb; </td>
<td> 27 February 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock down newbus infrastructure to support driver
fine-graining. </td>
<td> &a.imp; </td>
<td> 28 February 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Remove the MP safe syscall flag from the syscall table
and add explicit mtx_lock/unlock's of Giant to all
syscalls. </td>
<td> &a.dillon;, &a.mux;, &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> SMPng architecture document. </td>
<td> &a.jhb;, &a.rwatson; </td>
<td> 28 February 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Move to shared lock for VOP_GETATTR() to reduce blocking during
frequent lightweight VFS operations. Modify namei() to provide
a LOOKUP_SHARED flag to indicate when the lock required may be
shared instead of exclusive. </td>
<td> &a.jeff; </td>
<td> 11 March 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Document existing vm_map locking and verify it's correctness. </td>
<td> &a.alc; </td>
<td> 18 May 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Document existing vm_object locking and verify it's
correctness. </td>
<td> &a.alc; </td>
<td> 4 May 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Implement lazy interrupt thread switching (context
stealing) on i386. </td>
<td> &a.bmilekic;, &a.kan; </td>
<td> 10 December 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Implement lazy interrupt thread switching (context
stealing) on sparc64. </td>
<td> &a.jake; </td>
<td> 10 December 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Switch from using lockmgr in VM to using a mutex or exclusive
sxlock. Push down Giant on all VM except for vm_object/VFS and
vm_page/pmap components. </td>
<td> &a.alc; </td>
<td> 10 December 2002 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Modify device driver API to permit drivers to more easily
split "in interrupt context" and "in interrupt thread"
code so as to acknowledge interrupts faster. This will permit
lower latency in interrupt handling. </td>
<td> &a.peter;, &a.scottl; </td>
<td> 01 July 2005 </td>
<td> &status.wip; </td>
</tr>
<!-- Stalled -->
<tr>
<td> Make printf() safe to call in almost any situation to avoid
deadlocks. </td>
<td> &a.cp; </td>
<td> 15 May 2001 </td>
<td> &status.stalled; </td>
</tr>
<!-- Not Started -->
<tr>
<td> Remove priority argument from tsleep(), msleep(),
cv_*wait*(). </td>
<td> &nbsp; </td>
<td> 12 January 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Reimplement kqueue using condition variables. </td>
<td> &nbsp; </td>
<td> 15 March 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Conditionalize atomic ops in the SMP code that are used
for debugging statistics. </td>
<td> &a.peter; </td>
<td> 15 March 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Add a new witness check for exiting processes to verify that
an exiting process holds no locks. </td>
<td> &a.jhb; </td>
<td> 13 June 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Specify priorities for condition variables, semaphores, and
sx locks. </td>
<td> &nbsp; </td>
<td> 7 September 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Axe schedcpu() in favor of event driven priority updates as
much as possible. </td>
<td> &nbsp; </td>
<td> 7 September 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Fix *hold (e.g. crhold) to return reference to object. </td>
<td> &nbsp; </td>
<td> 7 September 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Add witness checking for lockmgr locks. </td>
<td> &nbsp; </td>
<td> 7 September 2001 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Add ICU spin locks on ia64. </td>
<td> &nbsp; </td>
<td> 4 January 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Fast-path push-down of Giant for VOP_READ() and VOP_WRITE(). </td>
<td> &nbsp; </td>
<td> 25 February 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Push the grabbing of Giant into Linux i386 ABI system calls.</td>
<td> &nbsp; </td>
<td> 25 February 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Push the grabbing of Giant into Linux AXP ABI system calls.</td>
<td> &nbsp; </td>
<td> 25 February 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Push the grabbing of Giant into SVR4 i386 ABI system calls.</td>
<td> &nbsp; </td>
<td> 25 February 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Push the grabbing of Giant into OSF/1 AXP ABI system calls.</td>
<td> &nbsp; </td>
<td> 25 February 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Push the grabbing of Giant into IBCS i386 ABI system calls.</td>
<td> &nbsp; </td>
<td> 25 February 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Expand mutex profiling tool to also profile sx locks. </td>
<td> &a.eivind;, &a.des; </td>
<td> 1 April 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Implement atomic_fetchadd() for int's and long's with acq and rel
versions. </td>
<td> &nbsp; </td>
<td> 23 May 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Implement a simple reference count API using atomic operations and
use this to replace locks that just protect a reference count. </td>
<td> &nbsp; </td>
<td> 23 May 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Add a witness_sleep() check to copyin/out() and s/fuword(). </td>
<td> &a.jhb; </td>
<td> 7 June 2002 </td>
<td> &status.new; </td>
</tr>
<!--
<tr>
<td> </td>
<td> &nbsp; </td>
<td> D M 2002 </td>
<td> &status.new; </td>
</tr>
-->
</table>
<p>This table lists the todo subtasks for multithreading the network
stack.</p>
<table class="tblbasic">
<tr class="heading">
<th> Task </th>
<th> Responsible </th>
<th> Last updated </th>
<th> Status </th>
</tr>
<tr>
<td> Protect network interface queues. </td>
<td> &a.jlemon; </td>
<td> 24 November 2000 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up IP. </td>
<td> &a.jennifer;, &a.hsu; </td>
<td> 10 June 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up TCP. </td>
<td> &a.jennifer;, &a.hsu;, &a.sam;, &a.rwatson;</td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up UDP. </td>
<td> &a.jennifer;, &a.hsu;, &a.rwatson; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock ifaddr reference counts. </td>
<td> &a.hsu; </td>
<td> 18 December 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up ifnet list. </td>
<td> &a.hsu; </td>
<td> 21 December 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock radix trees. </td>
<td> &a.hsu; </td>
<td> 23 December 2002 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up ARP. </td>
<td> &a.hsu; </td>
<td> 16 January 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up raw IP. </td>
<td> &a.sam;, &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock divert sockets. </td>
<td> &a.sam; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock ipfw2. </td>
<td> &a.sam; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock DUMMYNET. </td>
<td> &a.sam; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock ethernet bridge. </td>
<td> &a.sam; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock IP fragment queues. </td>
<td> &a.rwatson; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock routing entries. </td>
<td> &a.sam; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock FAST_IPSEC. </td>
<td> &a.sam; </td>
<td> 4 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Permit parallel entry into isr processing. </td>
<td> &a.rwatson;, &a.sam; </td>
<td> 11 October 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock if_disc "discard interface". </td>
<td> &a.rwatson; </td>
<td> 9 March 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock if_faith "IPv6-to-IPv4 TCP relay interface. " </td>
<td> &a.sam;, &a.rwatson; </td>
<td> 9 March 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock if_gif "generic tunnel interface". </td>
<td> &a.rwatson; </td>
<td> 9 March 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Review ECN tunnel support (ip_ecn.c). </td>
<td> &a.rwatson; </td>
<td> 9 March 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> if_tap global and softc locking. </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> if_tun global and softc locking. </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> netatalk/aarp.c locking. </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Cache socket MAC label in inpcb label for IPv4 sockets so that
the label can be used safely at the inet layer without socket
locks. </tD>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> IP encapsulation subroutines (ip_encap.c). </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock globals in loopback interface (if_loop.c). </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Use m_tags in if_gif to limit looping configurations, rather
than a non-MPSAFE static coutner. </td>
<td> &a.ru; </td>
<td> 23 April 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> netatalk DDP PCB locking. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock up syncache. </td>
<td> &a.hsu;, &a.sam; </td>
<td> 10 November 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Permit IP forwarding path to run Giant-free. </td>
<td> &a.sam; </td>
<td> 1 December 2003 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock &unix; domain protocols, fifofs. </td>
<td> &a.sam;, &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Giant lock over NFS server to protect against so_upcall() w/o Giant </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock interface cloning meta-data. </td>
<td> &a.brooks; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Apply combination of socket and socket buffer locks, label caching
to MAC labels on sockets so that they can be used safely without
Giant. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Make routing socket message dispatch use a netisr to avoid
re-entering the socket code from the routing code, resolving lock
order issues. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Introduce accept locking to protect accept incomplete and
complete queues on listen sockets. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Break out socket buffer wakeup, socket buffer append, socket
state change, socket buffer reserve, flush, etc, calls into
_locked() and unlocked versions, and avoid conditional locking.
</td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down AARP, AppleTalk Address Resoluton Protocol. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Fix pull/push cache data synchronization issues in sosend(),
soreceive(), allowing them to run Giant-free. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Protect socket global counters/limits and generation number
with a mutex. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down unit allocation meta-data in interface related
netgraph modules. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock down socket buffer OOB fields across TCP/IP, IPX. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add MSG_NBIO so that fifofs can avoid frobbing SO_NBIO in
a manner that risks races. </td>
<td> &a.truckman; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Protect all use of so_count with socket lock. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Move socket buffer related state from so_state to sb_state so
it can be properly locked by the socket buffer mutex. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Introduce a temporary global lock to lock the if_label field
used by the MAC Framework. </tD>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Push VFS-specific behavior out of fdrop_locked() and acquire
Giant in the fo_close per-object methods rather than
fdrop_locked(), so that pipes and sockets can run fo_close()
Giant-free. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Push Giant acquisition into fo_stat() file descriptor stat
operation, rather than acquiring it in fstat(), so that fstat()
on sockets and pipes can run Giant-free. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Don't hold socket locks over entry to protocol switch methods,
allowing protocol methods to acquire socket locks after protocol
locks in the lock order. </td>
<td> &a.rwatson; </td>
<td> 24 July 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Port inpcb mutex locking, assertions from IPv4 to IPv6. </td>
<td> &a.rwatson; </td>
<td> 8 August 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add IFF_NEEDSGIANT to allow if_start to run with Giant for
specific interfaces. Defer if_start to task queue. </td>
<td> &a.rwatson; </td>
<td> 8 August 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Push down Giant in stat(), fo_stat() to allow Giant-free stat
of pipes, sockets. </td>
<td> &a.rwatson; </td>
<td> 8 August 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Add TCP lock assertions. </td>
<td> &a.rwatson; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Lock socket layer. </td>
<td> &a.sam;, &a.rwatson; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Review TCP timer code. </td>
<td> &a.rwatson; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Analyze and reduce cost of entropy gathering in network critical
paths. </td>
<td> &a.rwatson;, &a.markm; </td>
<td> 24 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Allow code to declare NET_NEEDS_GIANT(), forcing Giant over the
network stack if that code is compiled into the kernel. </td>
<td> &a.rwatson; </td>
<td> 28 August 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Disable Giant over the network stack in the default
configuration. </td>
<td> &a.rwatson; </td>
<td> 28 August 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Additional KTR tracing for UMA, callouts, interrupts, etc. </td>
<td> &a.rwatson; </td>
<td> 07 November 2004 </td>
<td> &status.done; </td>
</tr>
<tr>
<td> Move to using file descriptor reference counts instead of
socket reference counts for socket system calls, avoiding extra
reference couht operations. </td>
<td> &a.rwatson; </td>
<td> 24 October 2004 </td>
<td> &status.done; </td>
</tr>
<!-- WIP -->
<tr>
<td> Lock IPv6. </td>
<td> &a.sam;, &a.rwatson;, &a.ume;, &a.mlaier; </td>
<td> 8 August 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> if_ppp global, per-softc locking. </td>
<td> &a.rwatson;, Maurycy Pawlowski-Wieronski &lt;maurycy@fouk.org&gt; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock struct ifnet. </td>
<td> &a.mlaier;, &a.luigi;, Maurycy Pawlowski-Wieronski &lt;maurycy@fouk.org&gt; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock IPv4, IPv6, atalk interface address lists. </td>
<td> &a.mlaier;, &a.rwatson; </td>
<td> 8 August 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock consumers of BSD compress (bsd_comp.c) code to protect
compression state. </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock global and softc state for six-to-four converter (if_stf.c). </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock down global and softc state for SLIP (if_sl.c). </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock global and softc state for SPPP (if_sppsubr.c). </td>
<td> &a.rik;, &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> IGMP locking. </td>
<td> &a.rwatson; </td>
<td> 23 April 2004 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> IP ID locking. </td>
<td> &a.ups; </td>
<td> 24 June 2005 </td>
<td> &status.wip; </td>
</tr>
<tr>
<td> Lock down netnatm. </td>
<td> &a.rwatson; </td>
<td> 01 July 2005 </td>
<td> &status.wip; </td>
<!-- Not Started -->
<tr>
<td> Reduce contention upon locking a socket buffer by replacing
tsleep() and wakeup() with a condvar. </td>
<td> &a.tanimura; </td>
<td> 21 April 2002 </td>
<td> &status.new; </td>
</tr>
<tr>
<td> Lock if_ef "ethernet frame" driver. </td>
<td> &nbsp; </td>
<td> 9 March 2004 </td>
<td> &status.new; </td>
</tr>
</table>
<a name="issues"></a>
<h2>Known Issues</h2>
<table class="tblbasic">
<tr class="heading">
<th> Issue </th>
<th> Last updated </th>
<th> Status </th>
</tr>
<tr>
<td> Idle processor time is not charged to the idle processes. </td>
<td> 20 September 2000 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> microuptime creeps backwards. </td>
<td> 4 October 2000 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> microuptime() went backwards </td>
<td> 4 October 2000 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> Process accounting is not accurate (the more CPUs, the
closer to correct it is). </td>
<td> 5 October 2000 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> M_DEVBUF is probably the wrong memory pool for interrupt
stuff and we should think about creating a new malloc pool for
that stuff. </td>
<td> 9 February 2001 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> PC card eject panics due to a race condition in the
interrupt thread code. </td>
<td> 15 March 2001 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> SMP x86 boxes are seeing NCPU * 100 clk interrupts and
NCPU * 128 rtc interrupts. </td>
<td> 15 May 2001 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> Witness will infinitely recurse when it acquires Giant after
sleeping with a sleepable lock. </td>
<td> 27 June 2001 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> Serial gdb does not work if boot_ddb and boot_gdb options
are specified. </td>
<td> 14 July 2002 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> Serial gdb does not work at 115200 baud. </td>
<td> 14 July 2002 </td>
<td> &status.resolved; </td>
</tr>
<tr>
<td> Serial gdb never regains control once 'cont' has been
entered. </td>
<td> 14 July 2002 </td>
<td> &status.resolved; </td>
</tr>
<!-- unresolved -->
<tr>
<td> Profiling is broken. </td>
<td> 20 February 2001 </td>
<td> &status.unresolved; </td>
</tr>
<!--
<tr>
<td> </td>
<td> D M 2002 </td>
<td> &status.unresolved; </td>
</tr>
-->
</table>
<a name="news"></a>
<h2>News</h2>
<p>The remainder of this page is structured as a
reverse-chronological log.</p>
<!--
<b> D M 2002</b>
<ul>
<li> <p> </p> </li>
</ul>
-->
<b>28 August 2004</b>
<ul>
<li>
&a.rwatson; threw the switch to change the network stack to run
without the Giant lock by default, permitting the network stack to be
run on multiple CPUs at a time, as well as to preempt and be
preempted by other code.
</li>
</ul>
<b>30 October 2003</b>
<ul>
<li>
<p>&a.grog; submitted a
<a href="http://www.lemis.com/grog/SMPng/Singapore/">FreeBSD SMPng paper</a>
to the <a href="http://www.aeosc.org/">Asian Enterprise Open Source
Conference</a> in Singapore. The paper presents a historical view of
SMPng development through 2001, but omits discussion of more recent
progress on the SMPng project, such as substantial performance enhancements
resulting from extensive lock pushdown in the storage subsystem, VM
subsystem, and major IPC subsystems.</p>
</li>
</ul>
<b>13 January 2002</b>
<ul>
<li>
<p>A <a href =
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=122646+0+archive/2002/freebsd-smp/20020113.freebsd-smp">
status report</a> was sent to the -smp mailing list.</p>
</li>
</ul>
<b>15 May 2001</b>
<ul>
<li>
<p>&a.grog; has made his <a href =
"http://www.lemis.com/~grog/SMPng/USENIX/">USENIX paper</a> available,
which he will present in Boston at the end of June.</p>
</li>
</ul>
<b>22 March 2001</b>
<ul>
<li>
<p>A <a href =
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=41302+0+archive/2001/freebsd-smp/20010325.freebsd-smp">
status report</a> was sent to the -smp mailing list.</p>
</li>
</ul>
<b>5 March 2001</b>
<ul>
<li>
<p>A <a href =
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=0+0+archive/2001/freebsd-smp/20010311.freebsd-smp">
status report</a> was sent to the -smp mailing list.</p>
</li>
</ul>
<b>24 January 2001</b>
<ul>
<li>
<p>A <a href =
"http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=38125+0+archive/2001/freebsd-smp/20010128.freebsd-smp">
status report</a> was sent to the -smp mailing list.</p>
</li>
</ul>
<b>12 January 2001</b>
<ul>
<li>
<p>A <a href =
"http://www.FreeBSD.org/cgi/getmsg.cgi?fetch=45111+48223+/usr/local/www/db/text/2001/freebsd-smp/20010114.freebsd-smp">
status report</a> was sent to the -smp mailing list.</p>
</li>
</ul>
<b>11 October 2000</b>
<ul>
<li>
<p>&a.jhb; and &a.cp; came up with a preliminary list of
<a href="http://storm.uk.FreeBSD.org/~hiten/smp_synch_rules.html">rules
</a> that should be followed when working on kernel
synchronization.</p>
</li>
</ul>
<b>6 September 2000</b>
<ul>
<li>
<p>The SMP code has been committed. All further work is being
done in cvs rather than with patches.</p>
</li>
<li>
<p>An updated <a
href="http://people.FreeBSD.org/~jasone/smp/smpng_patch_9.7">patch</a>
is available for download. This patch is probably what will
actually get committed.</p>
</li>
</ul>
<b>5 September 2000</b>
<ul>
<li>
<p>An updated <a
href="http://people.FreeBSD.org/~jasone/smp/smpng_patch_9.1">patch</a>
is available for download. This patch makes rtc a fast
interrupt, uses locked instructions for mutexes in MP kernels,
and corrects mtx_*() linkage within modules.</p>
</li>
</ul>
<b>1 September 2000</b>
<ul>
<li>
<p>The code is working for the most part now on i386 (UP and MP).
Some additional coding is still necessary for the alpha,
which is being done now.</p>
</li>
</ul>
<b>30 August 2000</b>
<ul>
<li>
<p>Updated patches for i386 and alpha are available <a href =
"http://people.FreeBSD.org/~jasone/smp/smpng_patch8.6">here</a>.</p>
</li>
</ul>
<b>12 August 2000</b>
<ul>
<li>
<p>Updated patches for i386 are available <a href =
"http://www.FreeBSD.org/~jhb/patches/smpng.patch6">here</a>.
Process accounting still doesn't work correctly, but a number
of other improvements have been made.</p>
</li>
</ul>
<b>3 August 2000</b>
<ul>
<li>
<p>Patches with functional heavy-weight threads for the i386
platform are available <a href =
"http://people.FreeBSD.org/~grog/patches4.gz">here</a>.
There are a couple of minor issues with this patch set.
Specifically, process accounting doesn't work correctly.</p>
</li>
</ul>
<b>6 July 2000</b>
<ul>
<li>
<p>Sheldon Hearn has prepared a mutex(9) man page based on the
BSD/OS one, which is available <a href =
"http://people.FreeBSD.org/~jake/mutex.9">here</a>.</p>
</li>
</ul>
<b>5 July 2000</b>
<ul>
<li>
<p>&a.jake; put an updated patch <a href =
"http://people.FreeBSD.org/~jake/smpng.diff">here</a>.</p>
</li>
</ul>
<b>3 July 2000</b>
<ul>
<li>
<p>&a.jake; has the BSD/OS lock code working now, and has
incorporated the pertinent portions of &a.dillon;'s patches
(idle processes, some of the schedlock changes, etc.).
His patch set is available <a href =
"http://people.FreeBSD.org/~jake/smpng2.tar">here</a>.</p>
</li>
</ul>
<b>26 June 2000</b>
<ul>
<li>
<p>&a.cp; has provided the PostScript versions of his
presentation slides for the
<a href="http://people.FreeBSD.org/~jasone/smp/smp_meeting_slides_1.ps">first day</a> and
<a href="http://people.FreeBSD.org/~jasone/smp/smp_meeting_slides_2.ps">second day</a> of the
SMP meeting.</p>
</li>
</ul>
<b>25 June 2000</b>
<ul>
<li>
<p>Here's a copy of the
<a href="http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=226208+0+archive/2000/freebsd-smp/20000625.freebsd-smp">SMP meeting summary</a>
that was posted to the -smp mailing list.</p>
</li>
</ul>
<b>19 June 2000</b>
<ul>
<li>
<p>Here's a copy of the
<a href="http://docs.FreeBSD.org/cgi/getmsg.cgi?fetch=154884+0+archive/2000/freebsd-current/20000625.freebsd-current">SMP project announcement</a>
that was posted to the -current mailing list.</p>
</li>
</ul>
&footer;
</body>
</html>
Reference in a new issue View git blame Copy permalink